Return to
Articles

Three Climate Modes

Because the oceans must abide by the constraints of geography and the laws of physics, there are only a few patterns in which the oceans can circulate. At a recent conference organized by Robert Webb of NOAA and Peter Clark of Oregon State University, climate scientists identified three modes of ocean circulation, each of which is associated with a different climate. The current mode produces the warmest conditions in the North Atlantic. Surface water sinks in two regions of the North Atlantic, and a large volume of surface water and heat is drawn from the tropics to replace the sinking North Atlantic water. The heat carried north by the northward-moving surface water warms eastern North America, the North Atlantic and most of Europe. The second mode of ocean circulation occurs when surface water sinks in only one area of the North Atlantic. Less surface water sinks to the bottom, so smaller amounts of warm surface water and heat are drawn north to replace the sinking water. This mode was in place during the warmest times of the Wisconsin ice age, when climate was only slightly colder than current conditions. In the coldest mode, no water sinks in the North Atlantic; hence no warm water is drawn north. This was the condition during the coldest portions of the Wisconsin ice age.

Each of these modes of ocean circulation is associated with a small range of prevailing environmental conditions. Weather anomalies such as 10-year-long droughts or wet periods, as significant as they may seem to human affairs, are a reflection of the small range of environmental conditions associated with a single mode of ocean circulation. If, however, environmental conditions are externally forced to be inconsistent with the existing mode of ocean circulation, the circulation will switch to a mode that is more consistent with those environmental conditions. An example of this forcing would be a change in the amount of solar heat reaching, and retained at, the earth’s surface. Such a change could be the result of alterations either in solar output or in the way the atmosphere regulates the exchange of heat between the earth’s surface and space. The transitions between different modes of ocean circulation are abrupt. The ocean-sediment cores and ice cores tell us that they frequently take only several decades or less.

Numerical models of ocean circulation developed by Thomas Stocker of the University of Bern and Syukuro Manabe of Princeton University show that each circulation mode is stable for a particular range of environmental conditions. For example, if the discharge of a river changes, altering the density of the surface water in the adjoining ocean, the ocean-circulation pattern will change only if it is unstable under this new set of conditions. As long as the climate system stays within the stable-mode range, river discharge and greenhouse-gas concentration can vary without having much influence on climate.

Stefan Rahmstorf of the Potsdam Institute for Climate Impact Research has used numerical models to show how surprisingly sensitive ocean circulation can be to changes in freshwater discharge. His numerical models show that if the climate system is near the threshold between stable modes, a small change in the amount of freshwater entering the North Atlantic will force a large and rapid shift to a different ocean-circulation pattern. Like a coin on edge, which topples with only a breath of air, an unstable pattern quickly assumes a new position where it becomes quite stable. The climate changes recorded by the ice and ocean-sediment cores appear to have taken place when some crucial threshold was crossed, resulting in large and rapid switches--in geologic time, like the flip of a switch--in ocean circulation.

Unfortunately, no one knows yet what caused this switch to flip. We know of external forcing mechanisms, but their time periods do not match the record. For example, the distribution of solar energy reaching the earth varies according to the relative positions of the sun and earth. These variations, called Milankovitch cycles, have periods of tens of thousands of years and are thus too slow to explain the rapid changes seen every couple of thousand years during the Wisconsin ice age. The Milankovitch cycles define the big picture and determine when changes could occur, but some smaller, quicker-acting mechanism triggered the more frequent switches during the Wisconsin.

The leading idea, which has been simulated in computer models, is that increased discharges of freshwater glacial ice and river runoff into the North Atlantic reduced the density of the surface water enough that it could not sink. This slowed the ocean conveyor, forcing it to switch to another circulation pattern. Other emerging concepts place the source of the disruption of the conveyor in the tropical Pacific. Variability in the sun’s output is another possible cause of the climate variations, but the record of solar output is not good enough to adequately investigate this idea. Furthermore, the dynamics of ocean circulation around Antarctica are too poorly understood to completely exclude the possibility that they may play a role. Whatever the cause may be, it is worrisome that the phenomenon that has repeatedly triggered major changes in ocean circulation and the earth’s climate is so subtle that we have not been able to identify it. This emphasizes how large changes in the interaction of the oceans, atmosphere and ice sheets have been triggered by small perturbations of the environment.

Tampering with Our Stable Mode?Human beings have made major modifications to the earth’s environment in little more than a century, increasing the concentration of carbon dioxide in the atmosphere to its highest level in 260,000 years. Numerical models can be used to estimate what will happen when anthropogenic increases in the atmospheric concentration of gases such as carbon dioxide and methane block heat from leaving the earth. The increased concentration of these gases acts like a greenhouse, and the average temperature of the earth gets warmer. But the numerical models of Stocker, Rahmstorf and others suggest there may be surprises in the greenhouse.

When the greenhouse effect warms the earth, it accelerates the hydrologic cycle, more water moves around in the atmosphere, and rainfall increases in many places. Some models suggest that this will slowly decrease the salinity of the North Atlantic, making the surface water less dense. Were a critical density threshold to be crossed, ocean circulation would abruptly switch to a new stable mode.

This would be more than just a switch in ocean circulation; it would be a switch in the way tropical heat is transported to the North Atlantic. At the least, Northern Europe and Scandinavia would be 2 to 5 degrees colder on average than they are now, and the amount of precipitation would decrease dramatically. It would not necessarily be a rapid return to an ice age, but it might be a start in that direction. The orbital parameters of the earth are such that we are due for another ice age, and a cooling in the north Atlantic at a time when orbital parameters favor a return to a much colder climate could be the trigger that initiates such a change.

A switch in climate from a warm period (like the current Holocene epoch) to an ice age has happened before. Ocean and lake cores tell us that the warm Eemian period from about 131,000 to 114,000 years ago--when the distribution of ice sheets was similar to what it is today--switched to the Wisconsin ice age in no more than 400 years, the minimum time resolution of the record from these ancient sediments. Unfortunately, we have yet to recover an ice core that shows in sharp detail how the Eemian Period ended. This is old ice. It is difficult to find a place where it snowed enough to produce a high time-resolution record but not so much as to smear the record against the bedrock. An international project, led by Claus Hammer of the University of Copenhagen, has identified the most likely location in Greenland for this ice to be found and is collecting a core.

Many arctic ice cores tell us that 8,200 years ago the climate approached ice age conditions for a 400-year period before returning to conditions similar to today. This excursion was most likely caused by the one-time draining of lakes left behind by the melting of the Canadian ice sheets. This change in freshwater flux to the oceans was large but not that much different from what greenhouse-induced changes may produce in the future. The fact that it took place when climate conditions were similar to today demonstrates that large and rapid climate switches do not happen exclusively when there are extensive northern hemisphere ice sheets. It is ironic that greenhouse warming may lead to rapid cooling in eastern North America, Europe and Scandinavia, and it is possible that altered ocean circulation could lead to much larger changes. We have no experience predicting climate switches between stable modes, so it would be wise to expect surprises.

Climate and ChoicesClimate is the result of the exchange of heat and mass between the land, ocean, atmosphere, ice sheets and space. As long as changes to the land, ocean, atmosphere and ice sheets stay below the thresholds I have just described, climate changes will happen slowly. But the climate will change rapidly if those thresholds are crossed. So rapidly that it would be impossible to rearrange agricultural practices quickly enough to avoid stressing world food supplies. So rapidly that many species would not be able to adapt, because their habitat, already greatly reduced by human activities, would be eradicated.

Human ingenuity would most likely allow us to adapt to a rapid change in climate, but we would pay a larger price than our civilization has ever known. Imagine the economic and social cost of moving, in a 20-year period, most of our agricultural activities 500 miles south of their current locations. Imagine the social cost and famine if agriculture could not be relocated quickly enough. Even a short-duration event such as the Dust Bowl years in the 1930s had a large influence on American society. The Little Ice Age, which caused major resettlement in Europe in the 15th and 16th centuries, is a more likely analogue of where we might be headed.

Some have proposed that we could counterbalance the greenhouse effect by manipulating the global exchanges of heat and mass. Methods that have been discussed include blocking the Strait of Gibraltar to change the salinity of the North Atlantic, using airplane-distributed particles or large orbiting sunshades to shade the earth, and fertilizing the ocean with iron to promote production of carbon dioxide-consuming biomass. But we have a poor record of managing even small ecosystems and lack a complete understanding of the ocean-atmosphere interactions that govern our climate. Intentionally manipulating climate would not only be costly and imprecise; it would also be impossible to benefit some regions without adversely effecting others. It would be a risky experiment on the only planet we can call home.

Although we do not know the critical level of greenhouse-gas concentration, we do know that reducing the rate of greenhouse emissions would help in two ways. First, the atmospheric concentration of greenhouse gases would increase more slowly. Second, numerical models by Thomas Stocker and Andreas Schmittner of the University of Bern and others predict that the climate threshold will occur at a higher concentration of greenhouse gases if the concentration of greenhouse gases increases slowly. Slowing the rate of greenhouse-gas emissions would buy us more time to understand the consequences of our actions and might allow us to increase greenhouse-gas concentrations to a higher level before reaching the critical threshold.

It is true that computer models are not perfect; they indicate general patterns. And we need to improve our understanding in many areas before the models can pinpoint thresholds. For example, our understanding of the details of ocean circulation is poor, and the physics of cloud formation and their influence on heat exchange is elusive. When we model previous switches in climate, we can compare the model to the results of real-world experiments recorded in ocean sediments and ice cores. But when we model the future, we have no empirical basis to judge the model’s accuracy. If we take no action until we are completely confident the models are correct, then the only use for the models will be to explain what happened. Our insistence on a tested model is part of the reason society is continuing to conduct the largest experiment ever done, the experiment of increasing the atmospheric concentration of greenhouse gases.

It will be another 20 years before the climate changes that are predicted to be associated with the greenhouse effect become large enough to be unambiguously differentiated from naturally occurring variations in climate. As a society we have the choice of ignoring the warning signs that our studies have uncovered or taking some action.

I think we should spend the next 20 years aggressively investigating our options. We should continue to focus on improving our ability to predict climate change. At the same time, we should test the technologies and polices we might need to reduce greenhouse-gas emissions, implementing them on a small scale where there would be minimal economic and social disruption. I am not alone among scientists in anticipating that 20 years from now our society may have to choose between disruptions associated with our current approach to energy use and disruptions associated with adopting an approach to energy use that produces fewer greenhouse gases. Procrastination will prevent making an informed decision and will increase the social and economic costs.

Previous Section

Next Section

© American Scientist 1999